- Poor Decisions Impacted by Bias in Software Development
- Organizational Culture Changes and Impacts on the Individual & Decision Making
- Using Data Analytics to Improve Heath Care
- Is Your Organization Using Evidence-Based Decision Making?
- Exploratory Data Analysis with Watson Analytics
- Simple Predictive Modeling with Watson Analytics
- Rapid Reporting and Visualization Development with Watson Analytics
Beyond Bias provides an excellent overview of the five different categories of biases, similarity, expedience, experience, distance, and safety. As someone who has worked in the software development field for a number of years, I have witnessed accounts of poor decision making and its ramifications within the organizations I have worked for. Specifically within software development I have witnessed poor decision making during proposal efforts, architectural decisions, and delivery.
Bidding on work, especially within the federal space, can be a time consuming and involved effort by many parties. On technical proposals, it is essential that the correct mix of people are part of the proposal team. I have observed how biases within the expedience domain can often come to play within business development. One project I was recently involved with in the last few years suffered from lack of technical input during the proposal phase, and availability and confirmation bias led the team to under bid the work. This had negative consequences on expectation of scope and delivery efforts later on. The situation could have been prevented by ensuring more technical expertise was placed within the team and those making the decisions affecting scope did an objective evidence.
Software architecture design is an area that quite often is critical within the project phase. Poor architectural decisions can often lead to adverse outcomes down the road. One project in particular I was involved with chose a specific content management system as the core backbone of the architecture. The problem arose because of a distance bias of temporal discounting. Implementing a content management system would provide for faster and easier startup, rather than a more lengthy and complex custom solution. The problem was the fundamental nature of this project did not require many of the capabilities of the chosen platform. In addition to this, development efforts became more complex because the team had to work around constraints enforced by this architectural decision. In the end, the project suffered greatly from this early decision and cost the organization greatly.
Delivery processes in software development range from weekly agile development efforts, quality assurance initiatives, and security. I have witnessed problems in the past with delivery due to processes that simply were not as effective as they could have been. This was often the case when the sunk costs bias where leadership was unwilling to effect changes because of the time, resources and money spent on current processes and a reluctance to alter course.
In all three of the above situations, better outcomes could have been achieved through checks and balances, more time spent on analyzing the problem, and involving other stakeholders to bring new perspectives into the conversation. Poor decision making will badly impact an organizations reputation, profitability, and sustainability. In all situations it is highly advisable to recognize the many different biases that can potentially lead to detrimental choices.